images images listlengths 6 6 | prompt stringclasses 1
value | target_json stringlengths 469 548 | env stringclasses 5
values | tier stringclasses 3
values | gt_return float32 406 16.5k | gt_norm_reward float32 0 1 | fell bool 2
classes | episode_id int32 0 19 |
|---|---|---|---|---|---|---|---|---|
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The humanoid executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | humanoid | expert | 10,482.272461 | 0.994648 | false | 0 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The humanoid executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | humanoid | expert | 10,381.254883 | 0.983424 | false | 1 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The humanoid executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | humanoid | expert | 10,456.286133 | 0.991761 | false | 2 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 915], "phase": "terminal", "command": "walk forward", "command_followed": false, "stability": "stumbling", "fall_risk": "high", "gait_quality": 0.1, "predicted_reward": 0.1, "anomaly": "robot clipped through ground late in rollout", "explanation": "The humanoid executes a expert-tier forward-loco... | humanoid | expert | 9,561.412109 | 0.892335 | true | 3 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 609], "phase": "terminal", "command": "walk forward", "command_followed": false, "stability": "stumbling", "fall_risk": "high", "gait_quality": 0.1, "predicted_reward": 0.1, "anomaly": "robot clipped through ground late in rollout", "explanation": "The humanoid executes a expert-tier forward-loco... | humanoid | expert | 6,084.369141 | 0.506013 | true | 4 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The walker2d executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | walker2d | expert | 6,892.936035 | 0.994325 | false | 0 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The walker2d executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | walker2d | expert | 6,929.955078 | 1 | false | 1 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The walker2d executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | walker2d | expert | 6,870.120117 | 0.990828 | false | 2 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The walker2d executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | walker2d | expert | 6,899.322266 | 0.995304 | false | 3 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The walker2d executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | walker2d | expert | 6,838.066406 | 0.985914 | false | 4 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "wobbling", "fall_risk": "medium", "gait_quality": 0.56, "predicted_reward": 0.56, "anomaly": "inconsistent gait, occasional stumbling", "explanation": "The walker2d executes a medium-tier forward-locomot... | walker2d | medium | 5,998.319336 | 0.857188 | false | 0 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "wobbling", "fall_risk": "medium", "gait_quality": 0.55, "predicted_reward": 0.56, "anomaly": "inconsistent gait, occasional stumbling", "explanation": "The walker2d executes a medium-tier forward-locomot... | walker2d | medium | 5,926.01123 | 0.846104 | false | 1 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "wobbling", "fall_risk": "medium", "gait_quality": 0.56, "predicted_reward": 0.56, "anomaly": "inconsistent gait, occasional stumbling", "explanation": "The walker2d executes a medium-tier forward-locomot... | walker2d | medium | 6,012.597168 | 0.859377 | false | 2 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "wobbling", "fall_risk": "medium", "gait_quality": 0.56, "predicted_reward": 0.57, "anomaly": "inconsistent gait, occasional stumbling", "explanation": "The walker2d executes a medium-tier forward-locomot... | walker2d | medium | 6,077.462402 | 0.86932 | false | 3 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "wobbling", "fall_risk": "medium", "gait_quality": 0.55, "predicted_reward": 0.56, "anomaly": "inconsistent gait, occasional stumbling", "explanation": "The walker2d executes a medium-tier forward-locomot... | walker2d | medium | 5,788.947266 | 0.825093 | false | 4 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.14, "predicted_reward": 0.17, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The walker2d executes a simple-tier forward-locomotio... | walker2d | simple | 4,197.978516 | 0.581211 | false | 0 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.14, "predicted_reward": 0.16, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The walker2d executes a simple-tier forward-locomotio... | walker2d | simple | 4,110.047852 | 0.567732 | false | 1 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.14, "predicted_reward": 0.17, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The walker2d executes a simple-tier forward-locomotio... | walker2d | simple | 4,203.091797 | 0.581995 | false | 2 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.14, "predicted_reward": 0.17, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The walker2d executes a simple-tier forward-locomotio... | walker2d | simple | 4,162.847656 | 0.575826 | false | 3 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 160], "phase": "terminal", "command": "walk forward", "command_followed": false, "stability": "stumbling", "fall_risk": "high", "gait_quality": 0.1, "predicted_reward": 0.1, "anomaly": "robot clipped through ground late in rollout", "explanation": "The walker2d executes a simple-tier forward-loco... | walker2d | simple | 406.441101 | 0 | true | 4 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "wobbling", "fall_risk": "medium", "gait_quality": 0.52, "predicted_reward": 0.54, "anomaly": "inconsistent gait, occasional stumbling", "explanation": "The humanoid executes a medium-tier forward-locomot... | humanoid | medium | 8,277.170898 | 0.749647 | false | 0 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 564], "phase": "terminal", "command": "walk forward", "command_followed": false, "stability": "stumbling", "fall_risk": "high", "gait_quality": 0.1, "predicted_reward": 0.1, "anomaly": "robot clipped through ground late in rollout", "explanation": "The humanoid executes a medium-tier forward-loco... | humanoid | medium | 4,345.364746 | 0.312799 | true | 1 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "wobbling", "fall_risk": "medium", "gait_quality": 0.52, "predicted_reward": 0.54, "anomaly": "inconsistent gait, occasional stumbling", "explanation": "The humanoid executes a medium-tier forward-locomot... | humanoid | medium | 8,230.31543 | 0.744441 | false | 2 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "wobbling", "fall_risk": "medium", "gait_quality": 0.52, "predicted_reward": 0.53, "anomaly": "inconsistent gait, occasional stumbling", "explanation": "The humanoid executes a medium-tier forward-locomot... | humanoid | medium | 8,171.11084 | 0.737863 | false | 3 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "wobbling", "fall_risk": "medium", "gait_quality": 0.52, "predicted_reward": 0.53, "anomaly": "inconsistent gait, occasional stumbling", "explanation": "The humanoid executes a medium-tier forward-locomot... | humanoid | medium | 8,121.254883 | 0.732324 | false | 4 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.12, "predicted_reward": 0.14, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The humanoid executes a simple-tier forward-locomotio... | humanoid | simple | 5,608.415527 | 0.453132 | false | 0 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.12, "predicted_reward": 0.14, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The humanoid executes a simple-tier forward-locomotio... | humanoid | simple | 5,714.849609 | 0.464957 | false | 1 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.12, "predicted_reward": 0.14, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The humanoid executes a simple-tier forward-locomotio... | humanoid | simple | 5,749.376465 | 0.468793 | false | 2 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.12, "predicted_reward": 0.14, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The humanoid executes a simple-tier forward-locomotio... | humanoid | simple | 5,679.61084 | 0.461042 | false | 3 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 825], "phase": "terminal", "command": "walk forward", "command_followed": false, "stability": "stumbling", "fall_risk": "high", "gait_quality": 0.1, "predicted_reward": 0.1, "anomaly": "robot clipped through ground late in rollout", "explanation": "The humanoid executes a simple-tier forward-loco... | humanoid | simple | 4,700.99707 | 0.352312 | true | 4 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The humanoid executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | humanoid | expert | 10,214.277344 | 0.964872 | false | 5 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The humanoid executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | humanoid | expert | 10,458.020508 | 0.991953 | false | 6 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The humanoid executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | humanoid | expert | 10,406.676758 | 0.986249 | false | 7 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The humanoid executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | humanoid | expert | 10,367.621094 | 0.981909 | false | 8 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The humanoid executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | humanoid | expert | 10,427.329102 | 0.988543 | false | 9 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The humanoid executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | humanoid | expert | 10,499.023438 | 0.996509 | false | 10 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The humanoid executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | humanoid | expert | 10,416.091797 | 0.987295 | false | 11 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 991], "phase": "terminal", "command": "walk forward", "command_followed": false, "stability": "stumbling", "fall_risk": "high", "gait_quality": 0.1, "predicted_reward": 0.1, "anomaly": "robot clipped through ground late in rollout", "explanation": "The humanoid executes a expert-tier forward-loco... | humanoid | expert | 10,387.796875 | 0.984151 | true | 12 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The humanoid executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | humanoid | expert | 10,412.650391 | 0.986912 | false | 13 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The humanoid executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | humanoid | expert | 10,522.782227 | 0.999149 | false | 14 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 200], "phase": "terminal", "command": "walk forward", "command_followed": false, "stability": "stumbling", "fall_risk": "high", "gait_quality": 0.1, "predicted_reward": 0.1, "anomaly": "robot clipped through ground late in rollout", "explanation": "The humanoid executes a expert-tier forward-loco... | humanoid | expert | 1,530.052979 | 0 | true | 15 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The humanoid executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | humanoid | expert | 10,418.069336 | 0.987514 | false | 16 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The humanoid executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | humanoid | expert | 10,375.930664 | 0.982833 | false | 17 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The humanoid executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | humanoid | expert | 10,468.371094 | 0.993103 | false | 18 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The humanoid executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | humanoid | expert | 10,530.443359 | 1 | false | 19 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 288], "phase": "terminal", "command": "walk forward", "command_followed": false, "stability": "stumbling", "fall_risk": "high", "gait_quality": 0.1, "predicted_reward": 0.1, "anomaly": "robot clipped through ground late in rollout", "explanation": "The humanoid executes a medium-tier forward-loco... | humanoid | medium | 2,132.470703 | 0.066932 | true | 5 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "wobbling", "fall_risk": "medium", "gait_quality": 0.52, "predicted_reward": 0.54, "anomaly": "inconsistent gait, occasional stumbling", "explanation": "The humanoid executes a medium-tier forward-locomot... | humanoid | medium | 8,248.371094 | 0.746447 | false | 6 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "wobbling", "fall_risk": "medium", "gait_quality": 0.52, "predicted_reward": 0.54, "anomaly": "inconsistent gait, occasional stumbling", "explanation": "The humanoid executes a medium-tier forward-locomot... | humanoid | medium | 8,238.447266 | 0.745345 | false | 7 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 334], "phase": "terminal", "command": "walk forward", "command_followed": false, "stability": "stumbling", "fall_risk": "high", "gait_quality": 0.1, "predicted_reward": 0.1, "anomaly": "robot clipped through ground late in rollout", "explanation": "The humanoid executes a medium-tier forward-loco... | humanoid | medium | 2,582.901367 | 0.116978 | true | 8 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "wobbling", "fall_risk": "medium", "gait_quality": 0.52, "predicted_reward": 0.54, "anomaly": "inconsistent gait, occasional stumbling", "explanation": "The humanoid executes a medium-tier forward-locomot... | humanoid | medium | 8,226.135742 | 0.743977 | false | 9 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "wobbling", "fall_risk": "medium", "gait_quality": 0.52, "predicted_reward": 0.53, "anomaly": "inconsistent gait, occasional stumbling", "explanation": "The humanoid executes a medium-tier forward-locomot... | humanoid | medium | 8,103.432617 | 0.730344 | false | 10 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "wobbling", "fall_risk": "medium", "gait_quality": 0.53, "predicted_reward": 0.54, "anomaly": "inconsistent gait, occasional stumbling", "explanation": "The humanoid executes a medium-tier forward-locomot... | humanoid | medium | 8,282.505859 | 0.75024 | false | 11 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "wobbling", "fall_risk": "medium", "gait_quality": 0.52, "predicted_reward": 0.54, "anomaly": "inconsistent gait, occasional stumbling", "explanation": "The humanoid executes a medium-tier forward-locomot... | humanoid | medium | 8,233.242188 | 0.744766 | false | 12 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "wobbling", "fall_risk": "medium", "gait_quality": 0.52, "predicted_reward": 0.54, "anomaly": "inconsistent gait, occasional stumbling", "explanation": "The humanoid executes a medium-tier forward-locomot... | humanoid | medium | 8,251.90332 | 0.74684 | false | 13 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 565], "phase": "terminal", "command": "walk forward", "command_followed": false, "stability": "stumbling", "fall_risk": "high", "gait_quality": 0.1, "predicted_reward": 0.1, "anomaly": "robot clipped through ground late in rollout", "explanation": "The humanoid executes a medium-tier forward-loco... | humanoid | medium | 4,528.504395 | 0.333147 | true | 14 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "wobbling", "fall_risk": "medium", "gait_quality": 0.52, "predicted_reward": 0.53, "anomaly": "inconsistent gait, occasional stumbling", "explanation": "The humanoid executes a medium-tier forward-locomot... | humanoid | medium | 8,089.247559 | 0.728768 | false | 15 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 634], "phase": "terminal", "command": "walk forward", "command_followed": false, "stability": "stumbling", "fall_risk": "high", "gait_quality": 0.1, "predicted_reward": 0.1, "anomaly": "robot clipped through ground late in rollout", "explanation": "The humanoid executes a medium-tier forward-loco... | humanoid | medium | 5,113.255859 | 0.398116 | true | 16 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "wobbling", "fall_risk": "medium", "gait_quality": 0.52, "predicted_reward": 0.53, "anomaly": "inconsistent gait, occasional stumbling", "explanation": "The humanoid executes a medium-tier forward-locomot... | humanoid | medium | 8,088.52002 | 0.728687 | false | 17 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "wobbling", "fall_risk": "medium", "gait_quality": 0.52, "predicted_reward": 0.53, "anomaly": "inconsistent gait, occasional stumbling", "explanation": "The humanoid executes a medium-tier forward-locomot... | humanoid | medium | 8,160.056152 | 0.736635 | false | 18 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "wobbling", "fall_risk": "medium", "gait_quality": 0.52, "predicted_reward": 0.54, "anomaly": "inconsistent gait, occasional stumbling", "explanation": "The humanoid executes a medium-tier forward-locomot... | humanoid | medium | 8,233.552734 | 0.744801 | false | 19 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.12, "predicted_reward": 0.14, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The humanoid executes a simple-tier forward-locomotio... | humanoid | simple | 5,679.644531 | 0.461046 | false | 5 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.12, "predicted_reward": 0.14, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The humanoid executes a simple-tier forward-locomotio... | humanoid | simple | 5,638.09668 | 0.45643 | false | 6 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.12, "predicted_reward": 0.14, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The humanoid executes a simple-tier forward-locomotio... | humanoid | simple | 5,659.094727 | 0.458762 | false | 7 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.12, "predicted_reward": 0.14, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The humanoid executes a simple-tier forward-locomotio... | humanoid | simple | 5,747.520508 | 0.468587 | false | 8 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.12, "predicted_reward": 0.14, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The humanoid executes a simple-tier forward-locomotio... | humanoid | simple | 5,611.374512 | 0.453461 | false | 9 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.12, "predicted_reward": 0.14, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The humanoid executes a simple-tier forward-locomotio... | humanoid | simple | 5,740.561523 | 0.467814 | false | 10 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.12, "predicted_reward": 0.14, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The humanoid executes a simple-tier forward-locomotio... | humanoid | simple | 5,642.789063 | 0.456951 | false | 11 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.12, "predicted_reward": 0.14, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The humanoid executes a simple-tier forward-locomotio... | humanoid | simple | 5,616.419434 | 0.454021 | false | 12 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.12, "predicted_reward": 0.15, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The humanoid executes a simple-tier forward-locomotio... | humanoid | simple | 5,834.609863 | 0.478263 | false | 13 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.12, "predicted_reward": 0.14, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The humanoid executes a simple-tier forward-locomotio... | humanoid | simple | 5,651.766113 | 0.457948 | false | 14 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.12, "predicted_reward": 0.14, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The humanoid executes a simple-tier forward-locomotio... | humanoid | simple | 5,698.486816 | 0.463139 | false | 15 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.12, "predicted_reward": 0.14, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The humanoid executes a simple-tier forward-locomotio... | humanoid | simple | 5,746.601074 | 0.468485 | false | 16 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.12, "predicted_reward": 0.14, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The humanoid executes a simple-tier forward-locomotio... | humanoid | simple | 5,586.773438 | 0.450727 | false | 17 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "unstable", "fall_risk": "high", "gait_quality": 0.12, "predicted_reward": 0.14, "anomaly": "poor locomotion quality, high fall risk", "explanation": "The humanoid executes a simple-tier forward-locomotio... | humanoid | simple | 5,763.509766 | 0.470364 | false | 18 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 492], "phase": "terminal", "command": "walk forward", "command_followed": false, "stability": "stumbling", "fall_risk": "high", "gait_quality": 0.1, "predicted_reward": 0.1, "anomaly": "robot clipped through ground late in rollout", "explanation": "The humanoid executes a simple-tier forward-loco... | humanoid | simple | 2,757.400635 | 0.136366 | true | 19 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The walker2d executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | walker2d | expert | 6,881.194824 | 0.992525 | false | 5 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The walker2d executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | walker2d | expert | 6,863.05127 | 0.989744 | false | 6 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The walker2d executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | walker2d | expert | 6,850.371094 | 0.9878 | false | 7 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The walker2d executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | walker2d | expert | 6,831.34082 | 0.984883 | false | 8 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The walker2d executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | walker2d | expert | 6,926.541016 | 0.999477 | false | 9 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The walker2d executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | walker2d | expert | 6,837.046387 | 0.985758 | false | 10 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The walker2d executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | walker2d | expert | 6,855.768066 | 0.988628 | false | 11 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The walker2d executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | walker2d | expert | 6,819.145996 | 0.983014 | false | 12 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The walker2d executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | walker2d | expert | 6,809.186523 | 0.981487 | false | 13 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The walker2d executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | walker2d | expert | 6,843.507813 | 0.986748 | false | 14 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The walker2d executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | walker2d | expert | 6,854.80957 | 0.988481 | false | 15 | |
You are RoboMind VLA, a vision-language reward model for robot locomotion. You are shown keyframes from a MuJoCo locomotion rollout. The robot was commanded to "walk forward". Analyze the rollout and respond with ONLY a JSON object with these exact keys: timestep_range, phase, command, command_followed, stability, fall... | {"timestep_range": [0, 1000], "phase": "walking", "command": "walk forward", "command_followed": true, "stability": "stable", "fall_risk": "low", "gait_quality": 1.0, "predicted_reward": 1.0, "anomaly": null, "explanation": "The walker2d executes a expert-tier forward-locomotion rollout over 1000 steps, accumulating re... | walker2d | expert | 6,822.952637 | 0.983597 | false | 16 |
End of preview. Expand in Data Studio
README.md exists but content is empty.
- Downloads last month
- 98